Panel Ordinary Kriging
Also known as: ordinary kriging for panel data, longitudinal ordinary kriging, repeated-measures spatial kriging, panel geostatistical interpolation
Panel Ordinary Kriging extends the classical geostatistical interpolation method — Ordinary Kriging — to panel (longitudinal) datasets where the same set of spatial locations is observed repeatedly over multiple time periods. It produces optimal linear unbiased predictions at unsampled locations for each time slice, accounting for spatial dependence while leveraging the temporal structure of the repeated observations.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use Panel Ordinary Kriging when you have repeated georeferenced measurements at the same set of locations across multiple time periods and need to interpolate attribute values to unsampled locations at each point in time — for example, environmental monitoring networks, longitudinal public health surveillance, or repeat soil-sampling campaigns. It is appropriate when the spatial mean is unknown but can be assumed constant within each time slice (the key ordinary kriging stationarity assumption). Do not use it when the spatial trend varies systematically across the study area (use Universal Kriging instead), when covariate data are available to improve prediction (use Co-Kriging or regression-kriging), or when the goal is inferential modeling of relationships rather than interpolation.
Strengths & limitations
- Delivers optimal linear unbiased predictions at unsampled locations under second-order stationarity, exploiting spatial autocorrelation fully.
- Provides spatially and temporally varying kriging variance, giving honest uncertainty bounds for each interpolated map.
- Does not require a known spatial mean: the ordinary kriging constraint enforces unbiasedness without estimating a global trend.
- Naturally handles irregular spatial sampling networks common in environmental and geospatial panel studies.
- Consistent application across all time periods allows comparison of interpolated surfaces through time.
- Assumes second-order stationarity (constant mean and covariance depending only on distance) within each time period — an assumption that may fail in heterogeneous landscapes.
- Variogram estimation requires enough spatial observations per time slice; sparse cross-sections yield unreliable variogram models.
- Computationally intensive for large panels: solving the kriging system scales with the number of conditioning points, and the cost multiplies over T time periods.
- Does not explicitly model temporal autocorrelation; treating each period independently discards information about how spatial patterns evolve over time.
- Sensitive to outliers and measurement errors, which can distort the empirical variogram and downstream predictions.
Frequently asked
How is Panel Ordinary Kriging different from Space-Time Kriging?
Panel Ordinary Kriging applies standard Ordinary Kriging independently (or with pooled variograms) to each time slice in a panel dataset. Space-Time Kriging uses a joint space-time covariance model that simultaneously accounts for spatial and temporal autocorrelation, which is more powerful but also more complex to specify and estimate.
Do I need the same spatial locations at every time period?
Ordinary Kriging operates within each cross-section separately, so locations can in principle differ across periods. However, if the spatial coverage changes substantially over time, variogram models may not be comparable across periods, complicating temporal comparison of the resulting surfaces.
How many observations per time period do I need to estimate a reliable variogram?
A common rule of thumb is at least 30-50 spatial observations per cross-section for stable variogram estimation. With fewer locations, consider pooling time periods or imposing a parametric variogram model with fixed shape parameters estimated from the full panel.
When should I use Universal Kriging instead?
Use Universal Kriging when the attribute shows a systematic spatial trend (e.g., elevation increasing toward a mountain range). Ordinary Kriging assumes the spatial mean is unknown but constant; when a trend is present, Ordinary Kriging predictions and variances are biased.
Can Panel Ordinary Kriging account for covariates?
Not directly. If covariate information is available (e.g., elevation, land use), Regression-Kriging or Co-Kriging are more appropriate: they combine a regression on covariates with kriging of the residuals, typically outperforming plain Ordinary Kriging when covariates are predictive.
Sources
- Cressie, N. A. C. (1993). Statistics for Spatial Data (revised ed.). Wiley-Interscience. ISBN: 978-0471002550
- Matheron, G. (1963). Principles of geostatistics. Economic Geology, 58(8), 1246-1266. DOI: 10.2113/gsecongeo.58.8.1246 ↗
How to cite this page
ScholarGate. (2026, June 3). Panel Ordinary Kriging. ScholarGate. https://scholargate.app/en/spatial-analysis/panel-ordinary-kriging
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Co-krigingSpatial analysis↔ compare
- Ordinary KrigingSpatial analysis↔ compare
- Panel KrigingSpatial analysis↔ compare
- Panel Spatial RegressionSpatial analysis↔ compare
- Spatial AutocorrelationSpatial analysis↔ compare
- Universal KrigingSpatial analysis↔ compare